Channel-wise autoregressive entropy models for image compression

A system, comprising:

• a data processing apparatus;
• and
• a non-transitory computer readable medium in data communication with the data processing apparatus and storing instructions executable by the data processing apparatus and that upon such execution cause the data processing apparatus to perform operations comprising:
• processing data using a first encoder neural network to generate a latent representation of the data;
• processing the latent representation of data, comprising:
• processing the latent representation of data by a first quantizer to generate a quantized latent representation of data;
• and
• processing the latent representation of the data using a second encoder neural network to generate a latent representation of an entropy model, wherein the entropy model is defined by one or more probability distribution parameters characterizing one or more code symbol probability distributions;
• processing the quantized latent representation of data into a plurality of slices of quantized latent representations of data, wherein each slice of quantized latent representation of data is different from each other slice of quantized latent representation of data, and the slices are arranged in an ordinal sequence;
• generating, by a hyperprior processing network and from the latent representation of the entropy model, hyperprior parameters representing the probability distribution of the entropy model and a compressed representation of the hyperprior parameters;
• for each slice in the ordinal sequence of slices:
• receiving, by a first slice processing network, the hyperprior parameters representing the probability distribution of the entropy model and the first slice of quantized latent representation of data, and generating, by the first slice processing network, a compressed representation of the first slice, and a first augmented slice that represents the first slice of quantized latent representation of data and a latent residual prediction that is a prediction of a residual encoding and decoding error based on the hyperprior parameters;
• for each slice subsequent to the first slice in the ordinal sequence of slices:
• receiving, by a respective subsequent slice processing network, the hyperprior parameters representing the probability distribution of the entropy model and each respective augmented slice generated by each prior respective subsequent slice processing network and the first slice processing network, and generating, by the respective subsequent slice processing network, a compressed representation of the respective slice, a respective subsequent augmented slice that represents the respective subsequent slice of quantized latent representation of data and a respective latent residual prediction that is a prediction of a residual encoding and decoding error based on the hyperprior parameters and each prior subsequent augmented slice;
• wherein a combination of the compressed representation of the first slice and each compressed representation of each respective slice form a compressed representation of the data.

A non-transitory computer readable medium storing instructions that, when executed by one or more data processing apparatus, cause the one or more data processing apparatus to perform operations comprising:

• processing data using a first encoder neural network to generate a latent representation of the data;
• processing the latent representation of data, comprising:
• processing the latent representation of data by a first quantizer to generate a quantized latent representation of data;
• and
• processing the latent representation of the data using a second encoder neural network to generate a latent representation of an entropy model, wherein the entropy model is defined by one or more probability distribution parameters characterizing one or more code symbol probability distributions;
• processing the quantized latent representation of data into a plurality of slices of quantized latent representations of data, wherein each slice of quantized latent representation of data is different from each other slice of quantized latent representation of data, and the slices are arranged in an ordinal sequence;
• generating, by a hyperprior processing network and from the latent representation of the entropy model, hyperprior parameters representing the probability distribution of the entropy model and a compressed representation of the hyperprior parameters;
• for each slice in the ordinal sequence of slices:
• receiving, by a first slice processing network, the hyperprior parameters representing the probability distribution of the entropy model and the first slice of quantized latent representation of data, and generating, by the first slice processing network, a compressed representation of the first slice, and a first augmented slice that represents the first slice of quantized latent representation of data and a latent residual prediction that is a prediction of a residual encoding and decoding error based on the hyperprior parameters;
• for each slice subsequent to the first slice in the ordinal sequence of slices:
• receiving, by a respective subsequent slice processing network, the hyperprior parameters representing the probability distribution of the entropy model and each respective augmented slice generated by each prior respective subsequent slice processing network and the first slice processing network, and generating, by the respective subsequent slice processing network, a compressed representation of the respective slice, a respective subsequent augmented slice that represents the respective subsequent slice of quantized latent representation of data and a respective latent residual prediction that is a prediction of a residual encoding and decoding error based on the hyperprior parameters and each prior subsequent augmented slice;
• wherein a combination of the compressed representation of the first slice and each compressed representation of each respective slice form a compressed representation of the data.